Volatile Anesthetics Depress Calcium sup 2+ Transients and Glutamate Release in Isolated Cerebral Synaptosomes

For each day's experiments, synaptosomes were prepared by a modification of the method of Nicholls [24]and Bowman et al. [9]Following protocols approved by the University of Virginia Animal Research Committee, 250-350-g Dunkin-Hartley guinea pigs were killed by cervical dislocation, or by cardiac excision after anesthesia with 0.1 g/kg pentobarbital given by intraperitoneal injection. Cerebral cortexes (approximately 1 g) were rapidly excised, cooled to 0 degree Celsius, homogenized in 9 ml 0.3 M sucrose, and centrifuged at 1,500g for 10 min. The pellet obtained was centrifuged at 1,500g for 10 min again after resuspension in approximately the original volume of 0.3 M sucrose. The combined supernatants were centrifuged at 9,000g for 20 min and the pellet (crude synaptosome fraction) was resuspended in 5 ml 0.3 M sucrose. Aliquots (2 ml) of this suspension were layered on 5 ml 0.8 M sucrose and centrifuged at 9,000g for 30 min. Particles dispersed in 0.8 M sucrose solution were diluted with an equal volume of ice-cold incubation medium containing (in millimolar concentrations) NaCl 125, KCl 5, NaHCO³5, Na²HPO⁴1.2, MgCl²1, glucose 10, and hydroxyethylpiperazine-ethane sulfonic acid 20, at pH 7.4 and centrifuged at 16,000g for 20 min. The pellet (purified synaptosome fraction) was resuspended in incubation medium.

Measurements of intracellular Calcium²+ and glutamate release were performed using a luminescence spectrofluorometer (LS-50, Perkin-Elmer Ltd., Beaconsfield, Buckinghamshire, England). During each experiment, the cuvette temperature was thermostatically regulated at 37 degrees Celsius with magnetic stirring of the synaptosome suspension to ensure adequate mixing.

The free Calcium²+ concentration inside the synaptosomes ([Ca sup 2+]ⁱ) was measured using the Calcium²+ -sensitive fluorophore Fura-2 and conventional excitation wavelength ratio methods. [9,25]The synaptosomal suspension was preincubated for 5 min at 37 degrees Celsius, followed by the addition of 100 micro Meter CaCl², 16 micro Meter bovine serum albumin, and 1 micro Meter Fura-2/AM (Sigma Chemical, St. Louis, MO), which diffuses into the synaptosomes and remains there after hydrolysis of the acetoxymethyl (AM) groups. After 35 min, the synaptosomes were centrifuged for 2 min and washed twice with 2 ml fresh medium to eliminate residual Fura-2/AM. An aliquot of Fura-2/AM-loaded synaptosomes (containing approximately 1.34 mg protein) was then placed in a cuvette containing 2 ml medium to which CaCl²was added to achieve a final [Calcium²+] of 1.3 mM. Following standardized protocols, fluorescence was determined at 510 nm with an excitation wavelength switched between 340 or 380 nm; data points were collected every 1.89 s. One hundred seconds after the addition of Calcium²+, [Potassium sup +] was increased to 35 mM (by addition of 20 micro liter 3 M KCl) to depolarize the synaptosomes and initiate Calcium²+ influx. After 180 s, synaptosomes were made soluble with 1% (weight/volume) Triton X-100 to permit Calcium²+ entry and saturation of Fura-2/AM, followed by addition of ethylene-glycol-bis-(beta-aminoethyl ether) tetraacetic acid to a final concentration 7.5 mM to bind Calcium²+, thereby generating the maximum and minimum 340/380 fluorescence ratios, respectively. Based on the maximum and minimum ratios, a computer program then calculated the synaptosomal [Ca²+]ⁱfrom fluorescence ratio measured during the 280-s experimental period. In control experiments, this technique was found to be insensitive to the anesthetics. Seven to 13 [Ca²+]ⁱassays on different synaptosomal preparations were carried out for each anesthetic concentration.

Glutamate release was measured under identical conditions used for Calcium²+ influx, except for omission of Fura-2/AM loading and the addition of an enzyme-coupled assay using glutamate dehydrogenase as described by Nicholls and coworkers. [7,26]In the presence of glutamate dehydrogenase (50 U/ml), nicotinamide adenine dinucleotide phosphate (NADPH) and ketoglutarate are produced from NADP sup + (1 mM) and glutamate. NADPH fluorescence was excited at 340 nm and monitored at 460 nm (emission wavelength) as a measure of glutamate release. As in the Calcium²+ assay, glutamate release was activated by increase in Potassium sup + to 35 mM 100 s after the addition of Calcium²+ to the solution. The addition of KCl increased the osmolarity of the solution by approximately 20%, but control experiments in which an equivalent amount of NaCl was added showed neither an increase in [Ca²+]ⁱnor detectable glutamate release.

Because of the enzyme-coupled aspect of this assay, it requires time for reaction. The sudden increase in glutamate due to addition or release results in a measurable fluorescence signal that increases exponentially to a final value. For a standard enzyme reaction, velocity of substrate utilization (-dS/dt, where S = substrate, in this case glutamate) is given by: Equation 1where P = the product and K^m= the substrate concentration at which 50% of the maximum velocity (V^max) is obtained for a constant enzyme concentration. Because the glutamate concentration is much lower than K^mof glutamate dehydrogenase (approximately 120 micro Meter), the equation may be simplified to: Equation 2Thus the initial rate of NADPH production is proportional to the glutamate concentration ([S]), and any sudden increase in glutamate can be estimated from the initial slope of NADPH appearance. To assess both the linearity of the assay as well as the validity of the assumption, the fluorescence intensity (FI) was measured in response to the sudden increase in glutamate caused by the addition of a specified amount (0.25-7 nmol) to the incubation medium, which contained 50 U/ml glutamate dehydrogenase and 1 mM NADP sup + (Figure 1(A)). As predicted from enzyme kinetics, the increase in FI was an exponential function of time after the addition of glutamate with a time constant of approximately 190 s at each glutamate concentration, the final FI stabilizing after 800-900 s. To 7 nmol added glutamate, the initial slope was a linear function of the added glutamate and the resulting [glutamate] (Figure 1(B)). Unlike the calibration curves, which achieve a stable FI once the added glutamate is metabolized, the initial and subsequent ongoing synaptosomal glutamate release activated by KCl depolarization results in a slow, ongoing increase in FI after 900 s. Consequently, absolute value of FI at any point cannot be used as a measure of initial or total glutamate release. However, because the initial slope of FI is a linear function of added glutamate to 7 nmol, the initial glutamate released can still be estimated from the initial slope. Although a deconvolution of the FI response based on the time constant of the reaction permitted an estimate of ongoing glutamate release, this was not routinely performed. To provide for any variation in each assay, the change in FI slope resulting from the addition of 6 nmol glutamate at 900 s was used an internal calibration for that particular assay (see Figure 4, for example). Five to eight glutamate assays on different synaptosomal preparations were carried out for each anesthetic concentration. We also verified that this method of glutamate determination is insensitive to the volatile anesthetics.

Anesthetic effects were determined by incubation of the synaptosomes for 5 min before study in an assay solution preequilibrated with the specified anesthetic concentration. Solutions were preequilibrated for 20 min by bubbling with anesthetic-containing filtered air from a calibrated vaporizer, whereas control solutions were bubbled with filtered air only. Incubation and assays were carried out in cuvettes in which an appropriate concentration of anesthetic-containing air continually flowed through the head space to prevent loss of anesthetic from the medium to the atmosphere. The Calcium²+ influx or glutamate release assay were performed with approximately equipotent anesthetic concentrations: 0.75 or 1.5% halothane; 1.3 or 2.5% isoflurane; 1.7 or 3.4% enflurane; representing one or two times human minimum alveolar concentration (MAC), or approximately 0.8 and approximately 1.6 times guinea pig MAC. [27]The aqueous phase anesthetic concentrations at 37 degrees Celsius as determined by gas chromatography are listed in Table 1and represent 89-105% of the value predicted from published partition coefficients. [28].

To decrease the amount of Calcium²+ entry and thereby assess the relation between the Calcium²+ transient and glutamate release, assays were also performed in the presence of 25, 50, 100, 200, 400, and 600 micro Meter Calcium²+, achieved by addition of smaller quantities of CaCl²to the synaptosome-containing assay solution before the KCl-induced activation.

Comparison among the absolute values of control and the two anesthetic concentrations and were performed using analysis of variance and the Fisher's protected least significant difference test for planned comparisons (Statview, Abacus Software, Berkeley, CA). Significance of the concentration dependence of the effects was determined by calculating the slope and its estimated error (least-squares regression analysis) using the absolute values at 0, 1, and 2 MAC of anesthetic. Values were also calculated as percent of control for any given days experiments (one to three assays), and tested versus 100% by an unpaired t test.

The increase in [Ca²+]ⁱas well as glutamate release from these isolated synaptosomes was clearly dependent on [Calcium²+] in the external medium ([Ca²+]^e). The depolarization induced by increasing [Potassium sup +] from 5 to 35 mM produced a characteristically rapid peak (< 10 s) of [Ca²+]ⁱof 100-150 nM above the basal concentration (120-200 nM), followed by a partial decrease to a stable plateau phase of 50-80 nM above the basal [Ca²+]ⁱ(Figure 2(A)). In nominally Calcium²+ -free solutions, there was never an increase in [Ca²+]ⁱ. In such cases, the contaminant [Ca²+] sub e in the absence of a Calcium²+ chelating agent was typically in the range of 2-4 micro Meter.

In the presence of external Calcium²+, the release of glutamate was evident from the increase in NADPH FI after KCl-induced depolarization, where the initial slope is proportional to the amount of the initial glutamate release (Figure 2(B)). The slope is subsequently maintained by ongoing glutamate release. Based on the FI slope resulting from addition of 6 nmol glutamate (tracing at 900 s not shown), the initial glutamate release was estimated as 1.5 nmol glutamate from the initial slope of the curve. In 30-40% of experiments, a modest baseline of glutamate release before depolarization was present as evidenced by a slightly positive slope of FI. In these cases, KCl-independent release of glutamate was subtracted from KCl-dependent release before tabulation. Depolarization in nominally Calcium²+ -free solution did not activate glutamate release, indicating that Calcium²+ influx is responsible for the observed increase.

If instead of 1.3 mM or nominally zero Calcium²+, an intermediate [Ca²+]^ewas used, [Ca²+]ⁱtransients of intermediate amplitude were obtained, and glutamate release was also reduced below the control (1.3 mM) concentration. Figure 3(A) presents the effects of 25-400 micro Meter [Ca²+]^eon the amplitude of the [Ca²+]ⁱtransient and plateau initiated by the Potassium sup + -induced depolarization. A modest reduction occurred with 400 micro Meter [Ca²+]^e, with smaller [Ca²+]ⁱtransients evident in the presence of the lower [Ca²+]^e. In spite of various complicating considerations, the decrease in the [Ca²+]ⁱtransient was a simple linear function of the log of [Ca²+]^e(Figure 3(B)). This relation would be anticipated if the [Ca²+]ⁱtransients were a linear function of inward Calcium²+ current amplitude, and assuming current can be approximated by a simple conductance model, current will be proportional to the Calcium²+ equilibrium potential, which is proportional to log ([Ca²+]^e/[Ca sup 2+]ⁱ). The glutamate release, as assessed by the initial slope of increasing FI, also decreased as [Ca²+]^ewas decreased (Figure 3(C)). Figure 3(D) presents the dose-dependent depression in glutamate release, as well as in the [Ca²+]ⁱtransient, for the reductions in [Ca²+]^e. With a reduction in [Ca²+]^e, glutamate release was typically more depressed than was the Ca²+ⁱtransient. Neither the basal [Ca²+]ⁱnor the basal glutamate release before depolarization was altered by the decreased [Ca²+]^e(data not shown).

The effect of isoflurane on KCl-induced [Ca²+]ⁱchanges and glutamate release are shown in Figure 4. Both the [Ca²+]ⁱtransients and the initial slope of the glutamate assay were reduced by isoflurane, with a greater effect present at the higher concentration. The effects of clinically comparable concentrations of halothane (0.75 and 1.5%), enflurane (1.7 and 3.4%) and isoflurane (1.3 and 2.5%) caused quantitatively similar decreases in the peak and plateau [Ca²+]ⁱresponse (Table 2) and in glutamate release (Table 3). Basal [Ca²+]ⁱand glutamate release were not altered by the anesthetics. The changes in [Ca²+]ⁱand secretion of glutamate induced by three anesthetics were also documented by the regression analysis. A significant negative slope of peak [Ca²+]ⁱtransient and initial glutamate release versus anesthetic concentration was present for all three anesthetics. The concentration dependence, when calculated for equivalent potency (1 or 2 MAC), was similar for all three agents: Calcium²+ transients were reduced 11-14% per MAC, and glutamate release was reduced about 20-25% per MAC.

For the various anesthetics and for [Ca²+]^eequal to 100-600 micro Meter, Figure 5plots the quantity of glutamate release as a fraction of control (Q^glut/Q^glut-control) versus the peak [Ca sup 2+]ⁱtransient, also expressed as a fraction of control ([Ca²+]ⁱ/[Ca²+]ⁱ-control), where the control values were those observed in 1.3 mM [Ca²+]^ein the absence of anesthetic. The mean percent same-day control values are plotted. If glutamate release is a simple linear function of [Ca²+]ⁱ, then the points should fall on the line of unity (n = 1). Instead, the fractional reduction in glutamate release is slightly greater than that for [Ca²+]ⁱ. Assuming a simple model for Calcium²+ -dependent glutamate release in which n Calcium²+ ions induce exocytosis, then: Equation 3where n = the cooperativity of the Calcium²+. In general, the points defined for reduced [Ca²+]^eas well as for the three anesthetics fall in the range of the lines defined by n = 1 and n = 2 (Figure 5).

Although homogenization of brain tissue destroys neuronal cell bodies, the membranes of nerve endings reseal into synaptosomes, small functional sacs that retain not only synaptic vesicles but also the ability to secrete neurotransmitter in response to a depolarizing stimulus. Although they of course do not reflect intact neuronal function mediated by axonally transmitted depolarizations, synaptosomes have been widely used to elucidate the mechanisms of excitation-secretion coupling. In the mammalian cerebrocortical synaptosomes, the predominant excitatory neurotransmitter is glutamate, which is released when Calcium²+ entry is activated. [6-8,29]We found that isoflurane, enflurane, and halothane significantly depressed the synaptosomal [Ca²+]ⁱincrease and release of glutamate evoked by a depolarizing concentration of Potassium sup + (35 mM). These effects on Calcium²+ are consistent with those reported by Kress et al. [30,31]in a variety of neuronal and some nonneuronal cells as well as with electrophysiologic evidence suggesting decreased neuronal glutamate release. [4,5]In the current study, solutions equilibrated with roughly equivalent clinical concentrations of isoflurane, enflurane, and halothane produced similar degrees of inhibition. As anticipated, the increase in synaptosomal Calcium²+ ([Ca²+]ⁱ) and the release of glutamate were both clearly dependent on [Ca²+]^eand could be proportionally reduced by decreasing [Ca²+]^e. Furthermore, effects on [Ca²+]ⁱand glutamate release caused by a reduction in [Ca²+]^eto 400 micro Meter (approximately 30% of the 1.3 mM control value) replicated the actions of the anesthetics. Because in nominally Calcium²+ -free solution, KCl depolarization by itself did not increase [Calcium²+], release of Calcium²+ from any intrasynaptosomal stores present did not appear to be occurring.

The observation that the decrease in the [Ca²+]ⁱtransient and greater decrease in glutamate release caused by the anesthetics can be closely duplicated by decreasing [Ca²+]^ehas important implications. Because the relation between the reduced [Ca²+]ⁱtransient and glutamate release observed in the presence of the anesthetics can be seen in their absence with decreased [Ca²+]^e, it is likely that the intrasynaptosomal mechanisms responsive to Calcium sup 2+, which ultimately progress to vesicle exocytosis, are not markedly altered by the anesthetics. If the synaptosomal proteins that bind Calcium sup 2+ and then foster vesicle fusion with the membrane were directly depressed (or enhanced) by anesthetics, then for a given [Ca²+]ⁱtransient, the glutamate release should be less than (or greater than) that observed when the transient was depressed by altering [Ca²+]^e. If either the anesthetics or decreased [Ca²+]^ealtered the resting [Ca²+]ⁱbefore depolarization, then the behavior of various Calcium²+ sensitive regulatory enzymes might influence subsequent exocytosis. However, no such change in [Ca²+]ⁱbehavior was noted. Consequently, the observed glutamate release from the synaptosomes can be largely explained by presuming the anesthetics mediate their action predominately by decreasing the abrupt increase in [Ca²+]ⁱresponsible for activating the exocytotic cascade. Whether this interpretation can be applied to intact neurons requires verification.

The reduction of [Ca²+]ⁱtransient is associated with a slightly greater decrease in the quantity of glutamate release (Q^glut/Q^glut-control). The suggestion has been made that the process of exocytosis shows cooperativity in its Calcium²+ dependence such that two or more Calcium²+ must bind to specific sites. [11,13,32,33]Such a dependence is typical of that observed previously, in which transmitter release is reduced in proportion to some power function of the reduction in [Calcium²+]. The current observations using decreased [Ca²+]^eare consistent with the presence of some degree of cooperativity (n > 1), and the anesthetics do not appear to markedly alter the degree of cooperativity.

Neurotransmitter release from neurons is mediated by Calcium sup 2+ entry into nerve terminals, activating a complex of proteins that cause fusion of the membrane of the transmitter-containing synaptic vesicle with the cell membrane, resulting in exocytosis. [23,34]Calcium²+ entry appears to be mediated by specific VGCC that are located near the active synaptosomal release zone of the neuronal membrane [32]and that are insensitive to the Calcium²+ -entry blockers such as the dihydropyridines classically active in the cardiovascular system. [35]Glutamate exocytosis appears to be coupled to Calcium²+ entry through several VGCC types including the N-, P- and Q-types, which are sensitive to omega-CTx-GVIA, omega-Aga IVA, and omega-CTx-MVIIC, respectively. [9,11,13-16,36,37]Although the exact VGCC may vary with the neuron, [38]Calcium²+ entry through VGCC appears to be of major importance because the depolarization-coupled glutamate release is much more efficient than that observed with the nonlocalized Calcium²+ entry due to the Calcium²+ ionophore ionomycin. [8,24,29]As with the L-type VGCC, volatile anesthetics appear to inhibit Calcium²+ currents in hippocampal neurons, [39]with a prominent effect observed on N-type and an unidentified, possibly P- or Q-type, VGCC. [21]In a separate investigation, P-type VGCC were interpreted as insensitive to volatile anesthetics, [22]although the authors' assumption that anesthetic potency (gas phase) is higher at room temperature may have underestimated the dose-requirement. Nevertheless, because of the cooperativity of Calcium²+ in mediating release, [11,13,32,33]even the 10% depression observed at approximately 1 MAC may result in more profound actions on presynaptic transmitter release.

Although association does not indicate causation, the results are consistent with a scheme in which inhibition of VGCC by volatile anesthetics accounts for inhibition of neurotransmitter release. However, anesthetic-mediated alterations in other aspects of cellular Calcium²+ regulation have been described that could contribute to the observed effects. [40,41]Volatile anesthetics have been shown to interfere with the sarcolemmal Calcium²+ -adenosine triphosphatase [42]as well as Sodium sup + -Calcium²+ exchange, [43]however, the role of these processes in regulating [Ca²+]ⁱand influencing neurotransmitter release in neuronal cells remains undefined. In synaptosomes, depolarization with [Potassium sup +] of 55 mM or less appears to activate Calcium entry only through VGCC, and does not activate the Sodium sup + -Calcium²+ exchange pathway. [44]Although Ca²+ⁱstores in neuronal endoplasmic reticulum or "calciosomes" may also contribute to changes in neuronal [Ca²+]ⁱ, [45]neurotransmitter vesicle release appears to be mediated by local domains of Calcium²+ that has entered near active zones at presynaptic endings. [46]In pheochromocytoma cells, anesthetics depress depolarization mediated Calcium²+ entry and norepinephrine release, but do not decrease Calcium²+ release from intracellular stores nor the norepinephrine exocytosis caused by receptor-activated synthesis of inositol trisphosphate. [31]In the current study, no significant change in resting [Ca²+]ⁱand glutamate release occurred in the presence of halothane, isoflurane or enflurane, suggesting that the anesthetics did not markedly alter the various Calcium²+ homeostatic pathways in resting synaptosomes. It is not possible to exclude indirect effects that might decrease the Calcium²+ transient, such as anesthetic depression of Sodium sup + influx on depolarization, which in turn could reduce Calcium²+ entry or enhance Calcium²+ elimination by the Sodium sup + -Calcium²+ exchanger. However, blockade of Sodium channels by tetrodotoxin does not inhibit glutamate release mediated by KCl-induced depolarization. [47].

The later phase of the NADPH FI assay contained components resulting from ongoing metabolism of the glutamate initially released and metabolism of glutamate released because of the ongoing analysis. Because these separate components could not be defined with certainty, the ongoing release was not quantitated. The plateau of the [Ca²+]ⁱwas significantly reduced by anesthetics or decreased [Ca²+]^e. However, the physiological relevance of the sustained depolarization and the associated ongoing glutamate release is unclear.

Although glutamate release stimulated by the KCl depolarization represents an artificial situation that imperfectly reflects in situ synaptic behavior, the anesthetic depression of glutamate release is consistent with the reported depression by halothane of the presynaptic glutamate release that generates excitatory postsynaptic potentials in thalamic [5]and hippocampal CA1 neurons. [48]It is unclear to what degree a 15-25% decrease in glutamate release, observed with 1 MAC anesthetic, could by itself interfere with the capacity of neurons to integrate and communicate information. However, when such an action on presynaptic endings is combined with the enhancement of gamma-aminobutyric acid^A-mediated inhibitory activity also caused by the volatile anesthetics, [1,2]the resulting effect should be more profound. Such combined actions might cause a greater alteration in behavior of individual neurons as well as in entire neural networks, and may also explain the differences in the quality of the anesthetic state as well as the neurophysiologic behavior produced by the volatile agents when compared with more pure gamma-aminobutyric acid^A-activating agents (barbiturates and benzodiazepines). [49].

The authors thank Joseph J. Pancrazio, Ph.D., for helpful discussion and suggestions throughout this project.